The world-wide growth of frequency spectrum efficient communications systems and the enhancement of their performance has increased the number of individual users and data transmission rates for these systems. Packet-based communication systems whose physical layers are based on multicarrier modulation are commonly referred to as OFDM (Orthogonal Frequency Division Multiplexing) or DMT (Discrete MultiTone) systems. The available transmission channel bandwidth within an OFDM system is subdivided into a number of discrete channels or carriers. Even though these channels overlap, they are orthogonal to each other. Data is transmitted in the form of symbols that have a predetermined duration and encompass some number of carrier frequencies. Systems in compliance with IEEE 802.11a and 802.11g wireless LAN standards are well-known examples of such systems.
The conventional structure of packets in a packet-based data transmission system comprises a preamble, a header, and a data payload. The preamble is typically used to estimate the channel impulse response, derive settings for the automatic gain control circuits, and perform carrier frequency offset estimation. It is also used for synchronization and other physical layer functions. The header is typically used for conveying information about variable physical layer parameters such as the size of the data payload and the type of modulation being employed for a particular packet.
Transmission properties, in general, and the impulse response, in particular, of wireless communication channels are time-varying statistical quantities. These variations in channel conditions are caused by several factors which include but are not limited to relative movement between the transmitter and receiver and movement of objects such as automobiles, people, portable office furniture, etc. in the vicinity of either the transmitter, receiver, or both. An example may include the use of a subscriber terminal within a wireless communication system in an automobile. The quality of the connection is extremely high since the subscriber terminal has a direct visual connection to the antenna of a base station within the system. Next, a truck moves in front of the automobile, blocking the direct visual connection to the antenna. Consequently, the quality of the wireless connection degrades, which is detected by the base station system in connection with channel estimation. Therefore, even when the transmit and receive antennas are both fixed spatially, the channel between them can still vary with time.
In such a wireless data transmission system, these variations in the channel response result in a corresponding variation in the short-term data rate that can be supported by the channel. Thus, the design of most data communication systems enable communication on the wireless channel at different data rates. When the signaltonoise ratio (SNR) is high for a particular channel, the channel quality is deemed to be good. Thereby, higher data rates are used on this specific channel. When the SNR is low, however, channel conditions are poor and, as a result, lower data rates are selected for the specific channel. “Noise” as defined here with respect to the SNR is understood to include receiver thermal noise as well as radio frequency interference (RFI) in the passband of the communication system. These varying data rates for each packet are selected either manually by the user or automatically by the system. In either case, some method and apparatus for estimating channel quality is required.
Channel quality metrics are used to estimate channel quality and thereby select the appropriate data rates. Most channel quality metrics are related to or estimated from the SNR measured at the receiver in an effort to set the data transmission rates of each channel. A conventional approach uses the average SNR over several packets to calculate the channel quality metric. As a result, either a large number of short packets or a small amount of long packets must be obtained to provide a reliable average SNR. Thus, in many wireless data communication systems, accurate measurement of the received SNR is difficult to achieve over a short period of time.
Moreover, to add to the complexity of the derivation of the SNR, within OFDM-based systems, there are multiple SNRs that correspond to multiple sub-channels within a channel. There are numerous approaches towards the combination or averaging of all the SNRs corresponding to each sub-channel to arrive at one quality metric for a particular channel. Some approaches use either the minimum value or the maximum value of each SNR in lieu of the arithmetic average of all SNRs within a given channel.
In an ideal situation, the channel impulse response remains constant for at least two packet-transmission durations. In this case, the channel is understood to have a coherence time that exceeds the duration of two packets. Accordingly, a receiving station may be able to estimate the channel quality from a received packet and select an appropriate transmission rate for its next transmission back to the station that sent the packet. This scenario, however, relies upon symmetry in the channel conditions between the transmitter and the receiver.
A known approach for estimating the SNR uses other metrics, such as the packet error rate (PER), where the PER is a function of the SNR. Typically, the operating PER in modern wireless data communication systems is 1%. Thus, a large amount of packets are necessary to measure the operating PER with reasonable accuracy. As such, this method for estimation can take a considerable amount of time.
Other approaches use the Viterbi decoder path metric to estimate the average SNR in IEEE 802.11-type OFDM physical layer. This estimation, however, requires an extensive amount of time to obtain reliable values of the SNR for each channel. Accordingly, the adaptation of data rates with respect to these SNR estimations is slow.
Thus, in order to increase the long-term average data transmission rates on communication systems whose channel impulse response is time-variant, it is necessary to be able to adapt data transmission rates more quickly with respect to changing channel conditions. This requires metrics for which reliable estimates can be obtained quickly. Since most modern packet-based wireless data communication systems support packets of varying length, it is desirable for the fidelity of the computed metrics to be largely independent of packet length. In the alternative, it is desirable for the fidelity to be guaranteed for the smallest expected packet.
The present invention is directed to overcoming, or at least reducing the effects of one or more of the problems set forth above.